Cloud Cost Optimisation: 5 Strategies That Actually Deliver

The average enterprise is wasting 32% of its cloud budget. Not because engineers are irresponsible, but because the default cloud configuration is optimised for flexibility and uptime — not cost. Here are the five levers we pull when a client asks us to reduce their cloud bill.

1. Right-size Compute First

Most workloads are running on instances 2–4× larger than they need. A week of CloudWatch/Azure Monitor data combined with a right-sizing analysis tool (AWS Compute Optimizer, Azure Advisor) typically reveals immediate savings of 15–25% with zero risk. This is always the first lever.

2. Reserved Instances and Savings Plans

On-demand pricing is a premium product for variable workloads. Any workload running more than 60% of the time should be covered by a Reserved Instance or Savings Plan commitment. A 1-year commitment delivers 30–40% savings; 3-year, 50–60%. Most organisations are dramatically under-committed.

3. Storage Tiering

Data that hasn't been accessed in 30 days costs the same as data accessed 1,000 times per day — unless you move it to a cheaper storage tier. S3 Intelligent-Tiering and similar services in Azure/GCP do this automatically for a small monitoring fee that is almost always worth it.

4. Delete Zombie Resources

Unattached volumes, unused load balancers, idle NAT gateways, forgotten test environments — every cloud account accumulates these. A quarterly zombie resource sweep is unglamorous but effective. We've found clients paying $8,000/month for resources serving zero traffic.

5. FinOps Culture, Not Just Tooling

The teams with the lowest cloud waste share costs back to the teams generating them. When a feature team can see that their poorly-optimised query is costing $3,000/month extra, they fix it. Tooling helps, but accountability is the real lever.